斎田 グレゴール

Victor Yocco highlights the need for new research methods in designing agentic AI systems that not only perform tasks autonomously but do so with a foundation of trust, consent, and accountability, marking a shift from traditional usability-focused approaches.

Victor Yocco highlights the need for new research methods in designing agentic AI systems that not only perform tasks autonomously but do so with a foundation of trust, consent, and accountability, marking a shift from traditional usability-focused approaches.

Victor Yocco highlights the need for new research methods in designing agentic AI systems that not only perform tasks autonomously but do so with a foundation of trust, consent, and accountability, marking a shift from traditional usability-focused approaches.

Victor Yocco highlights the need for new research methods in designing agentic AI systems that not only perform tasks autonomously but do so with a foundation of trust, consent, and accountability, marking a shift from traditional usability-focused approaches.

1

1件のコメント

Libraries are actively testing how #AI fits into archival practice. At Goldey-Beacom College, staff and students developed a review-centered approach where AI supports drafting and people retain decision-making authority. A new case study documents the workflow and the principles behind it. See how they did it: https://bit.ly/45U3Tyl

10

2件のコメント

Now on #LTIBlog Learn how to set up a #GenAI community of practice for #librarians and patrons to explore AI from Mason Libraries librarians Dr. Heidi Blackburn, Chris Magee, and Trevor Watkins https://ow.ly/Yu8v50WmlPe

1

AI-powered toys are coming to a toy store near you ... But these are NOT toys. They are AI systems designed to converse with children, simulate emotional responsiveness, and hold attention. Calling them 'educational doesn’t change what they are, child-facing chatbots wrapped in 'cuddly wide eyed' appeal branding. As an anthropologist, I find this deeply troubling. Childhood play is not about efficiency or constant stimulation. It is how we learn to tolerate boredom, negotiate conflict, read real emotions, and build agency. These are not flaws to be optimised away. They are the foundations of being human. AI systems do the opposite. They remove friction. They respond instantly. They adapt to please. They never get tired, distracted or emotionally unavailable. When we place that dynamic in front of a developing child, we quietly recalibrate what “relationship” means long before that child has the capacity to question it. We are introducing emotional machines into children’s lives without knowing the long-term cognitive or psychological consequences. We would never trial a new drug on children without years of testing. So are we happy to trial synthetic companionship on their developing brains? Some of these toys have already been shown to generate inappropriate, unsafe and even sexually explicit responses. That alone should have triggered a regulatory halt. It didn’t. And then there’s the data. These toys listen. They record. They transmit. Children cannot consent to this. Parents are rarely told clearly what data is collected, how long it’s kept, or where it goes. If this were happening in a workplace or public space, we’d call it surveillance. In a child’s bedroom, it’s marketed as innovation. If AI toys are genuinely safe, beneficial and ethical, then they should withstand rigorous independent testing, strict regulation and hard public scrutiny. Until then, normalising them is reckless. Childhood is not a sandbox for unregulated AI.

41

1件のコメント

ArchivesSpace is a standards-based open source archives information management system, designed by archivists and supported by archival repositories around the world. It supports many core functions for managing and providing access to archival materials, including analog, hybrid, and born-digital content. Click to learn more at: https://ow.ly/TrGe50WyLAM #ArchivesSpace #ASpace #archives #contentmanagement

7

Following our major announcement earlier this week, I am releasing the official physical studio documentation of "Whisper of the Known 1-1." Seeing the delicate, hand-finished lace by 洪慧 HONG Wai Art merged with the conceptual intelligence of Sophia the Robot (Hanson Robotics Limited) is incredible in full resolution. The physical manifest is fully realized. Next stop: VOLTA Art Fair Basel. 🇨🇭 Swipe through for the physical details, and head over to the HUMANARTY page to watch the full high-resolution studio video. 👇 The full Curator's Note and Newsletter release is linked in the comments below. #ContemporaryArt #VoltaBasel #SophiaTheRobot #HongWai #ArtMarket #HUMANARTY

4

3件のコメント

"Museum of the Moving Image and the Tezos Foundation expand their partnership with a new series of artist commissions exploring blockchain as a creative material, featuring installations on the Museum’s Herbert S. Schlosser Media Wall, events, and a public workshop track, the FA2 Fellowship.  From November 2025 through January 2027, five artist pairs will present works on the 50-foot Media Wall in the lobby. Each participating artist will also release a “production artifact” from their process, such as source code, sketches, or generative tools. Eight artifacts will be published on Tezos across the cycle and available for the public to collect at no cost, offering a parallel record of experimentation and development."   https://lnkd.in/eVJ9HQAc

1

Fragmented ecosystems and limited supply: Why China cannot break free from Nvidia hardware for AI Key Insight: China remains unable to break free from Nvidia hardware dominance in AI due to fragmented domestic ecosystems and limited supply, with Chinese companies still relying heavily on nearly one million Nvidia HGX H20 processors purchased in 2024 alone. Background: Despite China’s multi-year strategic push for AI hardware self-sufficiency and massive investments in domestic chip development, U.S. export restrictions and internal fragmentation have hindered progress. In 2025, China froze new Nvidia H20 GPU purchases amid security concerns, while top Chinese AI players like Alibaba, Baidu, Tencent, and DeepSeek continue to depend on Nvidia’s hardware and CUDA software ecosystem for AI workloads. Top Three Reasons This Matters: 1. Nvidia’s supply scale and mature CUDA software stack create a chicken-and-egg challenge for China’s AI industry, as domestic hardware lacks both volume and developer support—DeepSeek abandoned training on Huawei’s Ascend chips due to performance issues, reverting to Nvidia GPUs. 2. U.S. export controls and a 15% sales tax on Nvidia and AMD AI hardware sold to China have fragmented the global AI supply chain, reinforcing reliance on U.S.-based Nvidia chips while limiting China’s ability to build a truly independent AI ecosystem. 3. The lack of common hardware and software standards within China’s AI ecosystem leads to market fragmentation, compounding the difficulties for Chinese firms like Biren, Huawei, and Moore Threads to scale and compete with Nvidia’s integrated ecosystem dominance. Credit Tom’s hardware

13

1件のコメント

Researchers are at the heart of GRAPHIA Project. Their expertise, curiosity, and collaboration shape how the project grows and how it can truly serve the Social Sciences and Humanities community. Within the project, we’ve launched a series of short clips featuring researchers involved in GRAPHIA. In these conversations, they share their role in the project, what motivates their work, and how they believe GRAPHIA can impact research communities.   🎥 The first clip in the series features Georgia Kelly Rubidge from Abertay University: https://lnkd.in/dYttDdXn  Get to know the people helping build a more connected research landscape.   #GRAPHIA #OpenScience #SSH #KnowledgeGraphs #ResearchCommunity #ScholarlyCommunication #ResearchInfrastructure

20

Authorship is no longer assumed. It has to be proven. The C2PA standard introduces a new layer to design. Not visible styling, but verifiable origin. Every image, edit, and transformation can carry a traceable record. This shifts the role of graphic design. Design is no longer only about communication. It is about accountability. In a landscape shaped by generative AI, provenance becomes part of the system, not an afterthought. The brands that adopt it early will define what trust looks like next. Read the full article: https://lnkd.in/eUDfNTPi #BrandStrategy #ArtificialIntelligence #ContentProvenance #GraphicDesign #DigitalTrust #DesignSystems #BrandGovernance #EthicalAI #VisualCommunication #FormedSpace

2

What if a simple sketch could reveal how students truly understand Human–Computer Interaction? In Sketching Snapshots, Sarah Sterman, Miriam Sturdee, and Angela Miguel explore sketching as more than a design technique, using it as a reflective lens into how students make sense of complex, evolving concepts in HCI. Through pre- and post-semester workshops across HCI courses in the US and UK, the authors analyzed 68 student sketches alongside interviews with students and instructors. The result: sketching emerges as a powerful, low-barrier way to surface implicit assumptions, shifting mindsets, and diverse interpretations of what HCI means. One highlight from the study: students’ surprise and delight when they received their original sketches back at the end of the semester, a moment that made learning visible, personal, and tangible. Beyond assessment, this work shows how reflective sketching can enrich teaching, spark deeper classroom conversations, and reshape how we approach HCI education. ✨ Highly recommend trying “Sketching Snapshots” in your own classes. How do you invite reflection into your design or HCI teaching? Read more: https://lnkd.in/gj6ipzgE #ACMTOCHI

47

Libraries are navigating rapid change in research behavior, AI adoption and collection needs. Inside the March edition of the ProQuest Content Solutions: ➡️ New primary source content across ProQuest #DigitalCollections ➡️ Ebook Central Reader App updates that improve mobile reading ➡️ Teaching and research support for the U.S. at 250 ➡️ Webinars and blogs exploring academic‑grade AI in libraries Read the full edition: https://ow.ly/Ic5u50YBhMX

12

1件のコメント